Using binary alcohol dependence diagnoses (affected/unaffected) in genetic studies necessarily ignores information about the clinical and etiological heterogeneity of alcohol dependence. With growing recognition that clinical diagnoses are likely to be far removed from the genetic underpinnings, there has been increasing effort to perform genetic analyses in human samples using alternative phenotypes that may more closely reflect the biological basis of the disorder. In the field of alcohol research, more recent human association studies have branched out beyond clinical diagnoses and analyzed a number of different, more focused phenotypes related to alcohol use, for example, alcohol consumption patterns, craving, and physiological response to alcohol. However, most of this work has been conducted in a largely nonsystematic, exploratory manner, with little justification for the inclusion or exclusion of particular phenotypes across studies. This pilot project will conduct a systematic analysis of the alcohol dependence phenotype, and related aspects of alcohol use, to parse the phenotype into genetically informed components for use in genetic association analyses. Data from the Virginia Twin Study will be used, in which >9250 adult twins were assessed with comprehensive psychiatric interviews, to conduct multivariate analyses examining the genetic architecture across alcohol dependence criteria, measures of alcohol use/drinking patterns, and comorbid psychiatric disorders. In the same way that factor analysis provides information about the underlying phenotypic factor structure, the use of genetically informative twin data provides information about the underlying genetic factor structure across the variables. The emergent factor structure from the twin analyses can then be used to create genetic factor scores for subjects in the Irish Alcohol Dependence Study (Project 2) for use as secondary phenotypes in genetic association tests, to compare with results from dependence diagnoses and model organisms. Accordingly, this pilot aims to be a first step toward tackling the inherent phenotypic complexity associated with alcohol use and dependence and related traits. Of particular relevance to the VCU-ARC, we hypothesize that expanding the phenotypes analyzed in human studies will provide greater congruence with results from model organism studies where alcohol-related traits, rather than alcohol diagnoses per se, are modeled.